著者

佐藤 一法 吉野 純一 内田 健

出版者

サレジオ工業高等専門学校

雑誌

サレジオ工業高等専門学校研究紀要 (ISSN:18812538)

巻号頁・発行日

vol.36, pp.45-49, 2010

---

Recently, there are some important problems about how a distributed genetic algorithm is implemented on a multi-core processor or a many-core processor. This paper examines parallelization efficiency of a distributedgenetic algorithm which is implemented on a multi-core processor using OpenMP. Our computational result shows that parallelizing the fitness calculation of a distributed genetic algorithm makes best efficiency to shorten elapsed time of a distributed genetic algorithm. However, the result also indicates that decoding genesin the fitness calculation decreases the effect of parallelization of a distributed genetic algorithm.